Research On Mobile Robot Active Olfaction Based On Biological-Inspired Behavior

At present, robot active olfactory technology has important value of application for some fields, such as industry, security, military affairs etc. The key problem of robot performing active olfactory task is that accurate judgment of doubtful target and rapid confirmation of concrete position about odor source or doubtful target.The olfactory tracking behavior of biology is simple and available, to search the gas leak source, this paper research and develop the robot active olfactory strategy. This lays a foundation for the application of mobile robot to carry out various odor-related dangerous tasks.This paper covers the following original points:1. Based on kinesis response of flatworm and tropotaxis behavior, this paper presents a robot biological-inspired active olfaction search strategy, the strategy doesn't depend on variety ratio of concentration to control robot searching direction and step size, and not the absolute concentration of olfactory sensors, which can improve the searching efficiency and ability of adapting environment. The robot active olfactory strategy has three modules, bionic olfactory searching module, avoiding block module, and odor identification module. According to real-time information variety of sensors, the expert aptitudinal decision-making system can make right choices independently, and the robot executes relevant action module. The avoiding block module adopts fuzzy control algorithm to control robot's advance.2. This paper presents a robot bionic odor source identification strategy based on circle behaviors, inspired from the biology's circle behavior nearby the odor source, such as moth etc. The strategy can determinate whether the target is gas source depended on single olfactory information. The strategy has simple design, not be restricted by environmental condition, and its high effciency is proved by simulation and experiment.3. This paper present a robot Biological-inspired active olfaction Strategy based on hunt behavior of Wandering Albatross, The strategy integrates the information from olfaction, ultrasonic and vision sensors, including zigzag searching, upwind searching vision approach etc. According to real-time information variety of sensors, and depend on conflict restrain rules, the aptitudinal decision-making algorithm based active apperceive-action can make robot determinate independently to executes appropriate action. The experiment proves that the strategy can improve the searching efficiency and precision, and robot can active search and locate the odor source using the strategy in turbulent environment.4. This paper presents a robot bionic odor source identification strategy based on cross-model, inspired from the human's and animal's cross-model apperceive method. The strategy integrates the olfaction information, vision sensors, and distance information by Bayes method based on cross-model. The strategy can improve the identification, and reduce the miscarriage of justice and ambiguity of info. That the strategy can help robot identify accurately the block and odor source is proved by simulation and experiment.